Summary Accurate pathological diagnosis is crucial for optimal management of cancer patients. For the ~100 known central nervous system (CNS) tumour entities, standardization of the diagnostic process has been shown to be particularly challenging - with substantial inter-observer variability in the histopathological diagnosis of many tumour types. We herein present the development of a comprehensive approach for DNA methylation-based CNS tumour classification across all entities and age groups, and demonstrate its application in a routine diagnostic setting. We show that availability of this method may have substantial impact on diagnostic precision compared with standard methods, resulting in a change of diagnosis in up to 12% of prospective cases. For broader accessibility we have designed a free online classifier tool (www.molecularneuropathology.org) requiring no additional onsite data processing. Our results provide a blueprint for the generation of machine learning-based tumour classifiers across other cancer entities, with the potential to fundamentally transform tumour pathology.

Link to publication record in Explore Bristol Research PDF-document This is the author accepted manuscript (AAM). The final published version (version of record) is available online via Lancet at https://www.sciencedirect.com/science/article/pii/S1470204517301559?via%3Dihub . Please refer to any applicable terms of use of the publisher. University of Bristol -Explore Bristol Research General rightsThis document is made available in accordance with publisher policies. Please cite only the published version using the reference above. Full terms of use are available:

The development of the mammalian cerebellum is orchestrated by both cell-autonomous programs and inductive environmental influences. Here, we describe the main processes of cerebellar ontogenesis, highlighting the neurogenic strategies used by developing progenitors, the genetic programs involved in cell fate specification, the progressive changes of structural organization, and some of the better-known abnormalities associated with developmental disorders of the cerebellum.

Recently, we described a machine learning approach for classification of central nervous system tumors based on the analysis of genome-wide DNA methylation patterns [6]. Here, we report on DNA methylation-based central nervous system (CNS) tumor diagnostics conducted in our institution between the years 2015 and 2018. In this period, more than 1000 tumors from the neurosurgical departments in Heidelberg and Mannheim and more than 1000 tumors referred from external institutions were subjected to DNA methylation analysis for diagnostic purposes. We describe our current approach to the integrated diagnosis of CNS tumors with a focus on constellations with conflicts between morphological and molecular genetic findings. We further describe the benefit of integrating DNA copy-number alterations into diagnostic considerations and provide a catalog of copy-number changes for individual DNA methylation classes. We also point to several pitfalls accompanying the diagnostic implementation of DNA methylation profiling and give practical suggestions for recurring diagnostic scenarios.Electronic supplementary materialThe online version of this article (10.1007/s00401-018-1879-y) contains supplementary material, which is available to authorized users.

According to the 2016 World Health Organization Classification of Tumors of the Central Nervous System (2016 CNS WHO), IDH-mutant astrocytic gliomas comprised WHO grade II diffuse astrocytoma, IDH-mutant (AII), WHO grade III anaplastic astrocytoma, IDH-mutant (AAIII), and WHO grade IV glioblastoma, IDH-mutant (GBM). Notably, IDH gene status has been made the major criterion for classification while the manner of grading has remained unchanged: it is based on histological criteria that arose from studies which antedated knowledge of the importance of IDH status in diffuse astrocytic tumor prognostic assessment. Several studies have now demonstrated that the anticipated differences in survival between the newly defined AII and AAIII have lost their significance. In contrast, GBM still exhibits a significantly worse outcome than its lower grade IDH-mutant counterparts. To address the problem of establishing prognostically significant grading for IDH-mutant astrocytic gliomas in the IDH era, we undertook a comprehensive study that included assessment of histological and genetic approaches to prognosis in these tumors. A discovery cohort of 211 IDH-mutant astrocytic gliomas with an extended observation was subjected to histological review, image analysis, and DNA methylation studies. Tumor group-specific methylation profiles and copy number variation (CNV) profiles were established for all gliomas. Algorithms for automated CNV analysis were developed. All tumors exhibiting 1p/19q codeletion were excluded from the series. We developed algorithms for grading, based on molecular, morphological and clinical data. Performance of these algorithms was compared with that of WHO grading. Three independent cohorts of 108, 154 and 224 IDH-mutant astrocytic gliomas were used to validate this approach. In the discovery cohort several molecular and clinical parameters were of prognostic relevance. Most relevant for overall survival (OS) was CDKN2A/B homozygous deletion. Other parameters with major influence were necrosis and the total number of CNV. Proliferation as assessed by mitotic count, which is a key parameter in 2016 CNS WHO grading, was of only minor influence. Employing the parameters most relevant for OS in our discovery set, we developed two models for grading these tumors. These models performed significantly better than WHO grading in both the discovery and the validation sets. Our novel algorithms for grading IDH-mutant astrocytic gliomas overcome the challenges caused by introduction of IDH status into the WHO classification of diffuse astrocytic tumors. We propose that these revised approaches be used for grading of these tumors and incorporated into future WHO criteria.

Sarcomas are malignant soft tissue and bone tumours affecting adults, adolescents and children. They represent a morphologically heterogeneous class of tumours and some entities lack defining histopathological features. Therefore, the diagnosis of sarcomas is burdened with a high inter-observer variability and misclassification rate. Here, we demonstrate classification of soft tissue and bone tumours using a machine learning classifier algorithm based on array-generated DNA methylation data. This sarcoma classifier is trained using a dataset of 1077 methylation profiles from comprehensively pre-characterized cases comprising 62 tumour methylation classes constituting a broad range of soft tissue and bone sarcoma subtypes across the entire age spectrum. The performance is validated in a cohort of 428 sarcomatous tumours, of which 322 cases were classified by the sarcoma classifier. Our results demonstrate the potential of the DNA methylation-based sarcoma classification for research and future diagnostic applications.

German Cancer Consortium (DKTK), partner site Heidelberg, Germany.

Tumors with histological features of pilocytic astrocytoma (PA), but with increased mitotic activity and additional high-grade features (particularly microvascular proliferation and palisading necrosis) have often been designated anaplastic pilocytic astrocytomas. The status of these tumors as a separate entity has not yet been conclusively demonstrated and molecular features have only been partially characterized. We performed DNA methylation profiling of 102 histologically defined anaplastic pilocytic astrocytomas. T-distributed stochastic neighbor-embedding (t-SNE) and hierarchical clustering analysis of these 102 cases against 158 reference cases from 12 glioma reference classes revealed that a subset of 83 of these tumors share a common DNA methylation profile that is distinct from the reference classes. These 83 tumors were thus denominated DNA methylation class anaplastic astrocytoma with piloid features (MC AAP). The 19 remaining tumors were distributed amongst the reference classes, with additional testing confirming the molecular diagnosis in most cases. Median age of patients with MC AAP was 41.5 years. The most frequent localization was the posterior fossa (74%). Deletions of CDKN2A/B (66/83, 80%), MAPK pathway gene alterations (49/65, 75%, most frequently affecting NF1, followed by BRAF and FGFR1) and mutations of ATRX or loss of ATRX expression (33/74, 45%) were the most common molecular alterations. All tumors were IDH1/2 wildtype. The MGMT promoter was methylated in 38/83 tumors (45%). Outcome analysis confirmed an unfavorable clinical course in comparison to PA, but better than IDH wildtype glioblastoma. In conclusion, we show that a subset of histologically defined anaplastic pilocytic astrocytomas forms a separate DNA methylation cluster, harbors recurrent alterations in MAPK pathway genes in combination with alterations of CDKN2A/B and ATRX, affects patients who are on average older than those diagnosed with PA and has an intermediate clinical outcome.